Zinc speciation and isotopic exchangeability in soils

polluted with heavy metals

William Diesing, Sokrat Sinaj, Geraldine Sarret, Alain Manceau, T. Flura,

Paolo Demaria, A. Siegenthaler, Valérie Sappin-Didier, Emmanuel Frossard

To cite this version:

William Diesing, Sokrat Sinaj, Geraldine Sarret, Alain Manceau, T. Flura, et al.. Zinc specia-tion and isotopic exchangeability in soils polluted with heavy metals. European Journal of SoilScience, Wiley, 2008, 59, pp.716-729. .

HAL Id: hal-00311803

https://hal.archives-ouvertes.fr/hal-00311803

Submitted on 21 Aug 2008

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.

https://hal.archives-ouvertes.frhttps://hal.archives-ouvertes.fr/hal-00311803

1

Zn exchangeability in soils 1

2

Zinc speciation and isotopic exchangeability in soils polluted with 3

heavy metals 4

5

6

W.E. DIESING a, S. SINAJa, G. SARRETb, A. MANCEAUb, T. FLURAa, P. DEMARIAa, A. 7

SIEGENTHALERc, V. SAPPIN-DIDIERd & E. FROSSARDa 8

a Swiss Federal Institute of Technology (ETH Zürich), Institute of Plant Sciences, 9

Eschikon 33, CH-8315 Lindau, Switzerland; b Environmental Geochemistry Group, 10

LGIT, University of Grenoble and CNRS, BP 53, 38041, Grenoble Cedex 9, France; 11

c Swiss Federal Office for Agriculture (BLW), Mattenhofstrasse 5, 3003 Bern, 12

Switzerland; and d Institut National de la Recherche Agronomique (INRA), UMR 13

Transfert sol plante et cycle des éléments minéraux dans les écosystèmes cultivés, 14

BP 81, 33883 Villenave d’Ornon Cedex, France 15

16

Correspondence: E. Frossard, E-mail: emmanuel.frossard@ipw.agrl.ethz.ch 17

18

19

20

21

22

23

24

gsarretZone de texte European Journal of Soil Science, August 2008, 59, 716–729 doi: 10.1111/j.1365-2389.2008.01032.x

2

Summary 25

A correct characterization of heavy metal availability is a prerequisite for the 26

management of polluted soils. Our objective was to describe zinc (Zn) availability in 27

polluted soils by measuring the isotopic exchangeability of Zn in soil/solution (E 28

value) and in soil/plant systems (L value), by assessing the transfer of Zn and 65Zn in 29

the fractions of a six-step selective sequential extraction (SSE) in incubated soils 30

and by identifying Zn forms in soils using extended X-ray absorption fine structure 31

(EXAFS) spectroscopy. We distinguished 3 pools of exchangeable Zn: the pool of 32

Zn exchangeable within 1 minute which is observed in all soils, Zn exchangeable on 33

the medium term, and the slowly and not exchangeable Zn. The amount of Zn 34

present in the 2 first pools was similar to the L value measured with T. caerulescens. 35

The 3 first steps of the SSE solubilized the 1st pool and a fraction of the 2nd pool. 36

Most of the 2nd pool and a fraction of the 3rd pool were extracted in the 4th step of 37

the SSE, while the rest of the 3rd pool was extracted in the final steps of the SSE. 38

The EXAFS study conducted on two soils showed that more than half of the Zn was 39

present in species weakly bound to organic compounds and/or outer sphere 40

inorganic and organic complexes. Other species included strongly sorbed Zn species 41

and Zn species in crystalline minerals. The EXAFS study of selected SSE residues 42

showed that the specificity and the efficiency of the extractions depended on the 43

properties of the soil studied. 44

45

Introduction 46

An accurate assessment of heavy metal availability is essential for the proper 47

management of polluted soils. Young et al. (2006) reviewed the use of selective 48

sequential extractions (SSE) and isotope exchange (IE) techniques to characterize 49

3

heavy metals availability and speciation on the solid phase of soils. They concluded 50

that although sequential extractions are rather easy to implement, their results are 51

flawed because of the lack of specificity of the extractants for given metal species, 52

and/or because of the adsorption or precipitation of metals occurring during the 53

extraction. IE techniques conducted in soil/solution systems allow the quantification 54

of ions located on the solid phase of the soil that can exchange with the same ion 55

present in the soil solution within a given exchange time (E value) (Sinaj et al., 56

1999). Ayoub et al. (2003) and Sinaj et al. (2004) showed that E values measured 57

after a long IE time in acidic soils are identical to the amount of isotopically 58

exchangeable Zn measured in pot experiments with different plant species (L 59

values) demonstrating that Zn that is isotopically exchangeable within a time frame 60

relevant for plant growth is the main source of Zn for plant nutrition. In most studies 61

E values are only measured after a single time of exchange (for instance 24h; Young 62

et al., 2000; Degryse et al., 2003; Nolan et al., 2005) although according to Young 63

et al. (2006) the study of the kinetics of IE has a lot of potential for assessing Zn 64

forms and availability in soils. 65

To assess whether a SSE could deliver relevant results on cadmium (Cd) 66

availability, Ahnstrom and Parker (2001) carried out a sequential extraction on soils 67

that had been labeled with a stable isotope of Cd (111Cd). They measured in all 68

extracts the total Cd content and the abundance of 111Cd and then compared these 69

results with the amount of soil isotopically exchangeable Cd. They concluded that 70

no single fraction of the sequential extraction or a combination of fractions 71

corresponded to the size of the isotopically labile Cd pool. 72

Extended X-ray Absorption Fine Structure (EXAFS) spectroscopy is also 73

used to assess the forms of heavy metals in soils (Manceau et al., 2002). Sarret et al. 74

4

(2004) explained the high proportion of isotopically exchangeable Zn observed in a 75

polluted soil by its high concentration in octahedral Zn weakly bound to organic 76

compounds identified with EXAFS. The sensitivity of EXAFS spectroscopy for 77

exchangeable species (bound to organics or weakly sorbed on minerals) is however 78

relatively weak as compared to precipitated and crystalline phases. On the opposite, 79

chemical extractions allow a better quantification of easily mobilized species as 80

compared to more recalcitrant ones due to non-specific dissolution and possible 81

formation of new species. Therefore, the combination of IE methods, SSE and 82

EXAFS spectroscopy may provide a better picture of exchangeable and non-83

exchangeable metal species (Scheinost et al., 2002; Sarret et al., 2004). 84

We analyzed Zn exchangeability and speciation in six polluted soils with IE 85

methods, both in soil/water systems (E values) and in soil/plant systems (L values), 86

SSE and Zn K-edge EXAFS spectroscopy. Our purpose was to establish 87

relationships between the Zn pools determined by the various techniques, and 88

between these pools and soil parameters such as total Zn content and pH. Moreover, 89

the specificity of SSE steps for extracting real Zn chemical species was tested by 90

comparing the distribution of Zn species determined by EXAFS spectroscopy in the 91

soil and in selected extraction residues. To test whether the extraction steps released 92

Zn species with a specific exchangeability, soils were labeled with 65Zn before the 93

SSE, and the specific activity was measured in each residue. 94

95

Materials and methods 96

Soils 97

We studied six soils that had been polluted with heavy metals. The soil from the 98

Institut National de la Recherche Agronomique (INRA soil) was sampled from a 99

5

field experiment performed near Bordeaux, France (44° 51’ N 00° 32’ W). This soil 100

had received 100 t of municipal digested and dehydrated sewage sludge per hectare 101

every second year between 1974 and 1993. A description of the experiment can be 102

found in Weissenhorn et al. (1995). Two soils from the Institut für Umweltschutz 103

und Landwirtschaft (IUL soils) were collected in a field experiment conducted near 104

Bern, Switzerland (46° 55’ N 07° 25’ E). Aerobically digested and dehydrated 105

sewage sludge had been applied to the IUL SS soil, while pig slurry had been 106

applied to the IUL PS soil. The amendment-loading rate for both soils was 5 t ha-1 107

year-1 from 1976 to 1996. A description of this field experiment is given in 108

Siegenthaler et al. (1999). The three remaining soils were collected in the vicinity of 109

industrial metal smelting facilities located close to Dornach, Switzerland (47° 25’ N 110

07° 35’ E), and in Evin (50° 25’ N 03° 01’ E) and Mortagne (50° 30’ N 03° 27’ E), 111

France. The Dornach soil has accumulated Cd, Cu, Ni and Zn from the deposition of 112

about 700 t dust year-1 from brass smelting that began in 1895 and continued into 113

the 1980s before the installation of emission filters and scrubbers (Geiger et al., 114

1993). The smelting facility in Evin began operations in 1894 and grew to become 115

the largest Pb and Zn ore processing plant in Europe before its closure in 2003. Until 116

1970, the Evin facility emitted approximately 5 t smelter dust day-1 (LASIR, 2000). 117

The Mortagne soil is heavily polluted with metal dust and slag from a Pb and Zn 118

smelter in operation between 1906 and 1968 (Manceau et al., 2000). Thiry et al. 119

(2002) estimate that 15’000 t of metals have been dispersed over 25 hectares 120

surrounding the smelter. 121

Approximately 100 subsamples of the INRA, IUL SS and IUL PS soils were 122

collected at random intervals within the surface horizon (0-20 cm) to obtain a 123

representative sample. For the Evin and Dornach soils, samples were randomly 124

6

taken after having removed the litter layer to obtain the maximum contaminant 125

concentration in the soil. The sample site in Mortagne is the so-called metallicolous 126

meadow. The A horizon was characterized by a distinct layer (5 to 8 cm thick) 127

starting at a 15 to 25 cm depth in which smelter ash and tailings had been spread out 128

and buried at the time the smelter was closed. A representative sampling was 129

obtained by collecting approximately 100 random subsamples between 5 and 40 cm 130

to include the heavily polluted layer. The soil samples were well mixed, air-dried for 131

at least one week and passed through a 2-mm sieve. Remaining plant debris was 132

removed by hand prior to analysis. Relevant soil characteristics are listed in Table 1. 133

134

Isotopic exchange kinetics, compartmental analysis and determination of 135

isotopically exchangeable Zn 136

IE kinetic experiments were carried out using a 1:10 soil solution ratio and 2 mM 137

CaCl2 as described by Sinaj et al. (1999). After shaking the soil solution suspension 138

on an end-over-end shaker for three days, the samples were removed, placed on a 139

magnetic stirring plate and stirred at 300 rpm. The soil suspension samples were 140

spiked with 1.3 to 2.5 kBq of carrier-free 65Zn added as ZnCl2 (NEN Biosciences, 141

Boston, USA; specific activity 2.0 GBq mg-1 Zn). Aliquots of the soil suspension 142

filtered through a 0.2 μm porosity cellulose acetate membrane (Minisart, Sartorius) 143

were removed at 1, 3, 10, 30 and 60 minutes, and at 1, 7 and 14 days. After the first 144

60 minutes of IE the flasks were left on the bench and they were put back on the 145

stirring plate one hour before sampling at 1, 7 and 14 days so as to minimize the 146

dispersion of soil aggregates that would have been caused by a continuous stirring. 147

The concentration of Zn in the solution (CZn) was measured after 60 minutes, 1, 7 148

and 14 days by ion chromatography as proposed by Sinaj et al. (1999). This method 149

7

measures the oxalate complexable Zn in the filtered soil suspension which 150

approximates the concentration of the hydrated and weakly complexed Zn species 151

(Cardellicchio et al., 1999) present in the solution. 65Zn activity in the solution was 152

measured at all sampling times by β liquid scintillation detection (Packard 2500) at 153

an emission energy of 325 keV. We measured the β-counts using 1 ml of filtrate 154

with 5 ml of scintillation liquid (Packard Ultima Gold) and corrected them for 155

quenching effects. 156

The decrease of the fraction of radioactivity remaining in the solution (rt/R 157

where rt is the radioactivity remaining in the solution expressed in Bq after t 158

minutes, and R the total introduced radioactivity expressed in Bq) was analyzed in 159

each soil with a compartmental analysis to assess the number of Zn containing 160

compartments (a compartment is defined as an amount of material that acts as 161

though it is well-mixed and kinetically homogeneous, Cobelli et al. (2000)). We 162

proceeded as proposed by Cobelli et al. (2000). If we consider that for a given soil 163

an apparent isotopic equilibrium is reached before or at 14 days of exchange (i.e. 164

that the fraction of radioactivity remaining in the solution has reached a constant 165

value), we can subtract the fraction of radioactivity remaining at equilibrium from 166

the rt/R values measured at earlier times during the experiment and analyze the 167

resulting curve. If a finite number of compartments can be identified it is possible to 168

separate the obtained curve in a sum of exponential terms that are a function of 169

exchange time. The final equation describing the change of radioactivity in solution 170

with time can then be written as follows: 171

tCN

ii

t ieBARr ×−−

=∑ ×+=

1

1 (1) 172

8

Where rt/R is the fraction of 65Zn remaining in solution at the time of sampling, A, Bi 173

and Ci are constants, N is the total number of compartments. 174

The theory of the compartmental analysis states that when the system is 175

closed, at a steady-state for the element studied, and when the tracer (here 65Zn) is 176

introduced in one injection within a very short time, then the number of total 177

compartment (N) is equal to the number of exponential terms (N-1) plus 1 (Cobelli 178

et al., 2000). 179

This analysis suggests the presence in all soils of a compartment of Zn 180

exchangeable during the first minute of exchange while other compartments differed 181

from soil to soil in their time limits and/or in their total numbers. To simplify the 182

subsequent analysis of the results we decided to consider only three pools of Zn (a 183

pool being defined as a group of compartments, Cobelli et al., 2000) in the rest of 184

the paper: the pool of Zn exchangeable within 1 minute which is observed in all 185

soils (pool 1), the pool of Zn that is exchangeable between 1 minute and apparent 186

isotopic equilibrium (pool 2) and the pool of Zn that can not be exchanged or that 187

exchanges very slowly (pool 3). The amount of Zn isotopically exchangeable within 188

a given time (Et value, mg kg-1) is calculated using the following equation: 189

tZnt r

RCmvE ×⎥

⎦

⎤⎢⎣

⎡×⎟⎠⎞

⎜⎝⎛= (2) 190

Where v/m is the solution to soil ratio (l kg-1), CZn is the Zn concentration (mg l-1) in 191

solution and rt/R is the fraction of 65Zn remaining in solution at the time t of 192

sampling. The amount of Zn present in the pool 1 (Epool1) is calculated considering 193

an exchange time of 1 minute, the amount of Zn present in the pool 2 (Epool2) is the 194

difference between the amount of Zn that has been exchanged at apparent isotopic 195

equilibrium and the amount of Zn exchangeable within one minute. The amount of 196

9

Zn present in the pool 3 (Epool3) is calculated as the difference between soil total Zn 197

and the amount of Zn that has been exchanged at apparent isotopic equilibrium. 198

199

Pot experiment with Thlaspi caerulescens, calculation of the L value 200

The most readily exchangeable Zn pools were labeled with carrier-free 65Zn (NEN 201

Biosciences, Boston, USA; specific activity: 2.0 GBq mg-1 Zn). De-ionized water, 202

nutrient solution and 65Zn were mixed well in to the soils to bring the soil moisture 203

content to 50% water holding capacity (WHC), to provide basal nutrients and to 204

obtain an activity of 2.4 MBq kg-1 soil. The activity was raised to 4.7 MBq kg-1 soil 205

in the Dornach soil due its high Zn fixing capacity. The nutrient solution provided 206

120 mg K kg-1 dry soil as K2SO4 and KH2PO4, 30 mg Mg kg-1 dry soil as MgSO4, 207

140 mg N kg-1 dry soil as NH4NO3, 60 mg P kg-1 dry soil as KH2PO4 and 58 mg S 208

kg-1 dry soil as K2SO4 and MgSO4. Incubation of the soils for 40 days at 21° C 209

under aerobic conditions was performed to allow the 65Zn to label the most readily 210

exchangeable pools. Each pot was filled with 400 g dry mass soil and the water 211

content was raised to 75% water holding capacity (WHC) before sowing the seeds. 212

The pot experiment consisted of a randomized block design of four 213

replicates. Ten seeds of T. caerulescens (Ganges ecotype) were sown and after 214

germination (~14 days) the plants were thinned to four plants per pot. Pots were 215

watered daily with de-ionized water to maintain 75% soil WHC. The plants were 216

grown under a controlled environment of 16°C /8 h night and 20°C /16 h day, at 217

70% relative humidity and a light intensity of 280 μmol m-2s-1. 218

After 80 days growth, the plants were harvested by cutting shoots at the soil 219

surface. The plant aerial biomass was washed with de-ionized water, dried at 85° C 220

for 24 hours and the dry mass was measured. Plant Zn concentrations were obtained 221

10

by grinding the plant material with an agate ball mill and using a dry-ash digestion 222

method adapted from Chapman and Pratt (1961). The method was performed by 223

incinerating a 1 g sample at 500° C for 8 hours, dissolving the residual ashes with 2 224

ml of 5.8 M analytical grade HCl that was further diluted to 50 ml with de-ionized 225

water before element and isotope measurements. Zn concentrations were determined 226

by ICP-OES (Varian Liberty 220) and 65Zn was measured using high purity Ge 227

bore-hole gamma detector (EAWAG, Dübendorf, Switzerland). All 65Zn 228

measurements were corrected back to the date of soil labeling. 229

Calculation of the L value (mg Zn kg soil-1) was performed with the 230

following equation proposed by Smith (1981): 231

introduced

plant

seedplant

ZnZn

ZnZnL

65

65

)( −= (3) 232

where Znplant (mg Zn plant-1) is the amount of Zn in the aerial parts of the plant, 233

Znseed (mg Zn plant-1) is the amount of Zn in the seed, 65Znplant (Bq plant-1) is the 234

amount of 65Zn in the aerial parts of the plant, and 65Znintroduced the total amount of 235

65Zn introduced to the soil (Bq kg-1 soil). This equation provides the most 236

conservative calculation of L since it assumes that all the seed Zn was redistributed 237

to the aerial portions of the plant. The concentration of Zn present in the seeds 238

(ZnSeed) was 34.0 ng Zn plant-1 (standard error: 0.1 ng Zn plant -1). 239

240

Selective Sequential Extraction and Total Digestion 241

Before conducting the selective sequential extraction (SSE), soils were labeled with 242

65Zn and incubated for 20, 85 and 120 days at 21° C. At the time of labeling, a 243

carrier-free 65Zn solution (NEN Biosciences, Boston, USA; specific activity: 2.0 244

11

GBq mg-1 Zn) was added as ZnCl2 to de-ionized water and mixed well with the soil 245

samples at the rate of 22.5 to 86.0 MBq kg-1 soil. Soils were maintained at 50% 246

water holding capacity and well aerated during the incubation period. 247

The selective sequential extraction used in this study was a six-step 248

procedure (F1-F6) developed by Salbu et al. (1998) and modified using a 1:10 soil 249

to extractant ratio. A description of the analytical grade reagents, procedures and the 250

proposed binding mechanisms are provided in Table 2. The extraction procedure is 251

designed to extract metals in a step-wise fashion first from weak outer sphere bound 252

forms (F1-F3), then from tightly bound outer and inner sphere complexes (F4 and 253

F5) and finally from crystalline metal forms (F6). Prior to performing extractions, 254

the soil samples were ground to a fine powder to homogenize and increase the 255

surface area exposed to the extractants during the extraction process. For each 256

extraction step the samples were shaken on a horizontal shaker at 100 rpm for the 257

times listed in the procedure. Following each extraction, the samples were 258

centrifuged at 11’000 g for 30 minutes, the solution was filtered through a 0.45 µm 259

porosity filter and the filtrate was analyzed for Zn and 65Zn. Measurements for Zn 260

were performed on an ICP-OES (Varian Liberty 220) and 65Zn was measured using 261

high purity Ge bore-hole gamma detector (EAWAG, Dübendorf, Switzerland). All 262

65Zn measurements were corrected back to the date of soil labeling. 263

As a final step (F7), the F6 residue was placed in an open-vessel microwave 264

digester (MX 350 Prolabo, France) and the procedure of Lorentzen and Kingston 265

(1996) was used to extract the residual Zn. After digestion, the sample was cooled to 266

room temperature filtered with a 0.45 µm porosity filter prior to ICP-OES (Varian 267

Liberty 220) analysis. In addition, a single step total digestion was conducted on all 268

12

soils to obtain total Zn and 65Zn quantities as a comparison to the total quantities 269

recovered by the SSE. 270

The fraction of Zn that had undergone isotopic exchange in each fraction of 271

the SSE was assessed by calculating the specific activity (SA) of Zn in each fraction 272

normalized by the quantity of isotopes introduced during labeling (65Znintroduced Bq 273

kg-1 soil) and the total soil Zn (Zntotal mg Zn kg-1 soil) as shown in equation 4. 274

⎟⎟⎠

⎞⎜⎜⎝

⎛

⎟⎟⎠

⎞⎜⎜⎝

⎛

=

total

Fx

introduced

F

ZnZn

ZnZn

SA

X

65

65

(4) 275

In this equation the subscript Fx is the extraction number, 65ZnFx represents the 276

radioisotope concentration (Bq kg-1 soil) in this extract and ZnFx the total Zn 277

concentration (mg Zn kg-1 soil) in the same extract. 278

279

Statistics 280

All soil analyses were conducted in triplicate while the plant analyses were made 281

with four replicates. Mean values are presented with the standard errors. 282

“Statgraphics plus for Windows” was used for both linear and nonlinear regressions. 283

The standard error of estimate (SEE) and the coefficient of determination are given 284

for each regression. The validity of regressions was evaluated by comparing the 285

predicted and experimental values and by looking at the residuals. 286

287

Extended X-ray Absorption Fine Structure (EXAFS) Spectroscopy 288

13

Untreated reference samples of the Dornach and Mortagne soils and residues 289

obtained from the selective sequential extraction were air-dried at 35° C ground and 290

pressed into 5 mm diameter pellets for EXAFS analysis. 291

Experiments were conducted in 2003 at the European Synchrotron Radiation 292

Facility (ESRF, Grenoble, France) on beamline ID-26. The electron storage ring was 293

operating in 16 bunch mode at 6 GeV and current ranging from 70 to 90 mA. The 294

monochromator was a pair of Si(220) flat crystals. Spectra were collected in 295

fluorescence mode using a photo-diode detector and aluminum filters. For each 296

sample 10-20 scans of 40 minutes were averaged. Data extraction was done using 297

WinXAS (version 2.0, Ressler, 2000). 298

Data analysis was done by linear combination fits (LCFs) without principal 299

component analysis since this latter approach is not adapted to small sets of spectra. 300

The LCFs were conducted using a reference Zn K-edge reference spectra library 301

described previously (Manceau et al., 2003; Sarret et al., 2004). The maximum 302

number of components for the fit of the extraction residues and soil spectra was 303

limited to four since the precision of the method does not enable a reliable 304

quantification of more complex mixtures. From these four components, 305

concentrations of Zn species (mg Zn kg-1 soil) in each sample were calculated by 306

multiplying the percentage of each component by the total Zn concentration in the 307

sample. 308

309

Results and discussion 310

Zn concentration in dilute CaCl2 extracts (CZn) 311

We present the average CZn values measured during the IE kinetic experiments for 312

each soil (Table 3). The highest values were observed in the Evin and Mortagne 313

14

soils and the lowest values were observed in the IUL SS soil (Table 3). The 314

logarithm of CZn was highly significantly related to soil pH and to the logarithm of 315

the total soil Zn content of the 6 soils. 316

ln(CZn)=4.81-3.06pH +1.86ln(Zntotal), n=6, R2=0.97, SEE=0.67 (5) 317

This result confirms that CZn increases with Zn inputs and decreases when soil pH 318

increases as noted by Arias et al. (2005). The very low CZn values observed in the 319

IUL SS soil can be explained by its high amorphous iron oxide content (Table 1) 320

acting as a strong sorbent for Zn. The elevated amorphous iron oxide content of this 321

soil is related to the repeated additions of FeCl3 treated sewage sludge in this field 322

experiment. 323

324

Decrease of radioactivity in dilute CaCl2 extracts with time during the isotopic 325

exchange kinetic experiments 326

The radioactivity found in the solution at a given time divided by the total amount of 327

radioactivity added to the suspension (rt/R) decreased following the same pattern in 328

all samples (Figure 1). In most soils the 2 last rt/R values were very similar showing 329

that an apparent isotopic equilibrium had been reached after 14 days. Only in the 330

IUL PS and in the Evin soils was the last rt/R value lower than the previous one 331

indicating that the isotopic equilibrium might have not been reached after 14 days of 332

IE. 333

It was possible to fit the curves describing the decrease of rt/R with exchange 334

time by the equation 1 by a sum of two exponential terms and a constant for the 335

Dornach soil and by a sum of three exponential terms and a constant for the other 336

soils (data not shown). These results suggest that the 65Zn added in the solution 337

exchanged with Zn located in 3 compartments in the Dornach soil and in 4 338

15

compartments in the other soils. These compartments corresponded to the following 339

exchange times: 0-1 min, 1-30 min, 30 min-7 days and > 7 days in INRA and IUL 340

SS; 0-1 min, 1-10 min, 10 min to 14 days and >14 days in IUL PS; 0-1 min, 1-30 341

min, 30 min to 14 days and > 14 days in Evin; 0-1 min, 1 min to 1 day and > 1 day 342

in Dornach; and 0-1 min, 1-30 min, 30 min to 7 days and > 7 days in Mortagne. 343

Models with two exponential terms and a constant were tried for all soils but gave a 344

proper fit only for Dornach (results not shown). This analysis suggests the presence 345

in all soils of a compartment of Zn exchangeable during the first minute, while other 346

compartments differed from soil to soil. 347

This compartmental analysis has some limits and its results must be 348

interpreted with caution. More sampling points might have resulted in the 349

determination of more compartments (Fardeau, 1993). Besides, it is extremely 350

difficult to sample the suspension at exchange time shorter than 1 minute (Fardeau, 351

1993). The stochastic approach used by Sinaj et al. (1999) was tested with our soils. 352

This approach allowed modeling the changes of rt/R with time as well as the sum of 353

exponentials for all soils except for IUL PS where it led to negative values of 354

radioactivity as time tended towards the infinity (results not shown). Besides, this 355

stochastic approach does not allow distinguishing compartments of exchangeable 356

elements (Fardeau, 1993). 357

358

Calculation of the amount of isotopically exchangeable Zn (E values) 359

Because of the different number of compartments observed between soils we prefer 360

to summarize the information given by the compartmental analysis by considering 361

only three pools of exchangeable Zn for each soil: the amount of Zn exchangeable 362

within 1 minute which is observed in all soils (pool 1), the amount of Zn 363

16

exchangeable that is exchangeable on the medium term (pool 2), and the amount of 364

Zn that is very slowly or not exchangeable (pool 3). The pool 2 corresponds to the 365

fraction of soil Zn that is exchangeable between 1 minute and apparent isotopic 366

equilibrium, i.e. between 1 minute and 1 day in Dornach, between 1 minute and 7 367

days in INRA, IUL SS and Mortagne, and between 1 minute and 14 days in IUL PS 368

and Evin. The pool 3 corresponds to the amount of Zn that could not be isotopically 369

exchanged within 1 day in Dornach, within 7 days in INRA, IUL SS and Mortagne 370

and within 14 days in Evin and IUL PS. 371

The amounts of Zn present in the 1st, 2nd and 3rd pools (Epool1, Epool2 and 372

Epool3) are presented in Table 3. Soils polluted with organic amendments (INRA, 373

IUL SS, IUL PS) had between 3.5 and 11.9% of the total Zn in the pool 1 and 374

between 68.1 and 78.9% of total Zn in the pool 3. The Evin and Mortagne smelter-375

impacted soils had between 42.3 and 46.8% of the total Zn in the pool 1 and 376

between 39.1 and 39.9% of total Zn in the pool 3. The Dornach smelter-impacted 377

soil showed an intermediate result with 10.7% of Zn in the pool 1 and 67.5% in the 378

pool 3. Highly significant relationships were observed between the logarithm of the 379

Zn content of pool 1 (Epool1) and pool 2 (Epool2) and the logarithm of total Zn and pH 380

(equations 6 and 7). 381

ln(Epool1)=-0.51-1.03pH+1.67ln(Zntotal), n=6, R2=0.98, SEE=0.39 (6) 382

ln(Epool2)=-2.92+0.19pH+1.03ln(Zntotal), n=6, R2=0.98, SEE=0.23 (7) 383

These equations show that the amount of Zn present in the 1st pool increases with Zn 384

inputs and decreases with pH, while the amount of Zn present in the 2nd pool 385

increases with Zn inputs and pH. 386

387

Pot experiment with T. caerulescens, calculation of the L value 388

17

The results are presented in Table 4. The biomass production of T. caerulescens was 389

similar in all soils, but the Zn content in the plant increased with soil total Zn 390

content. The L values varied between 22.2 and 32.9% of the total soil Zn content in 391

the three soils that had been polluted by organic amendments and between 33.4 and 392

56.4% of the total soil Zn content in the three soils that had been polluted by smelter 393

emissions. The L values were numerically very similar to the sum of the Zn content 394

present in the 2 first pools of the isotope exchange kinetic analysis (equations 8 and 395

9). 396

ln(L)=0.13+0.97ln(Epool1+Epool2), n=6, R2=0.99, SEE=0.08 (8) 397

ln(Epool1+Epool2)=-0.12+1.03ln(L), n=6, R2=0.99, SEE=0.08 (9) 398

This result shows that pools 1 and 2 contain the soil Zn that can be accessed by T. 399

caerulescens through diffusion and desorption. 400

401

Selective sequential extraction of 65Zn and Zn from incubated soil 402

No significant shifts in 65Zn concentration between the extracted fractions were 403

noted for all soils between 20, 85 and 120 days of incubation, with the exception of 404

the Evin and Mortagne soils in which 65Zn concentrations significantly decreased in 405

F2 while the concentration of 65Zn increased in F3 and F4. This coincides with the 406

findings of Almås et al. (1999, 2000) in which measurable levels of 65Zn were found 407

among all fractions within 7 days of soil labeling. 408

The normalized specific activities observed for each fraction are presented in 409

Table 6. A monotonous decrease down to null activities was expected from the most 410

exchangeable (F1) to the residual (F7) fraction. However, the specific activities 411

values obtained in F1 were often lower than in F2. We suggest that the relatively 412

low Zn and 65Zn concentrations extracted by the water (F1) and their high variability 413

18

might explain the difficulties in calculating correct specific activities values. If we 414

do not consider F1, a monotonous decrease is observed from F2 to F7, except for the 415

F3 extraction for the Evin and Mortagne soils. The activity is very low in the F6 416

extract and close to 0 in the F7 residual, as expected. This 65Zn tracing of the SSE 417

confirms that the chosen extractants induced a progressive removal of Zn from 418

highly exchangeable to recalcitrant species. 419

The average concentrations of Zn in the different fractions of the SSE are 420

provided in Table 5. The total amount of Zn recovered from this sequential 421

extraction ranged between 90.9 and 116% of the total Zn content measured after 422

direct digestion. Between 41.0 and 49.6% of the total Zn was extracted in the 6th 423

step (F6) in INRA, IUL SS and IUL PS soils while in the Evin and Mortagne soils 424

between 45.4 and 53.6% of the Zn was extracted in the 2nd and 3rd steps (F2-F3 425

fractions). This predominance of exchangeable species is most likely due to the 426

dissolution of smelter-inherited primary minerals (franklinite, sphalerite, willemite) 427

and redistribution in the exchangeable fractions as described by various authors 428

(Juillot et al, 2003, Manceau et al., 2000, Roberts et al., 2002). 429

Comparison between the amounts of Zn recovered in the different fractions 430

of the SSE (Table 5) and the amount of isotopically exchangeable Zn (Table 3), 431

shows that the total amount of Zn extracted during the 3 first steps (ZnF1+F2+F3) is 432

slightly higher than the amount of very rapidly exchangeable Zn (Epool1) (equations 433

10 and 11). 434

ln(ZnF1+F2+F3)=0.78+0.92ln(Epool1), n=6, R2=0.97, SEE= 0.38 (10) 435

ln(Epool1)=-0.68+1.05ln(ZnF1+F2+F3), n=6, R2=0.97, SEE=0.40 (11) 436

This suggests that these 3 first steps have extracted the entire quantity of Zn 437

isotopically exchangeable within 1 minute, and that F3 extracted a fraction of the Zn 438

19

exchangeable on the medium term. Highly significant correlations were also found 439

between ln(ZnF5+F6+F7) and ln(Epool3) (equations 12 and 13). 440

ln(ZnF5+F6+F7)=1.20+0.73ln(Epool3), n=6, R2=0.87, SEE=0.35 (12) 441

ln(Epool3)=-0.65+1.19ln(ZnF5+F6+F7), n=6, R2=0.87, SEE=0.44 (13) 442

The lower amount of Zn recovered in F5+F6+F7 compared to Epool3 suggests that a 443

fraction of very slowly or not exchangeable Zn had already been extracted in F4. 444

Since F1+F2+F3 extracted the Zn present in the pool 1 and some of the Zn present 445

in pool 2 and F5+F6+F7 extracted a fraction of the pool 3, we conclude that the 4th 446

step of the SSE solubilized both moderately and slowly exchangeable forms of Zn, 447

i.e. Zn from pools 2 and 3. 448

449

Zn K-edge EXAFS spectroscopy 450

Zinc K-edge EXAFS analysis was conducted on the Dornach and Mortagne 451

untreated samples and on the F2 and F3 residues for the Mortagne soil, and the F3, 452

F4 and F5 residues for the Dornach soil. 453

Figure 2 shows the Zn K-edge EXAFS spectra for some reference 454

compounds used in the linear combination fits, including franklinite, Zn-sorbed 455

birnessite (Mn oxide), Zn-substituted kerolite as a proxy for Zn-substituted 456

phyllosilicate, Zn/Al hydrotalcite, a zinc-aluminum hydroxycarbonate, Zn-sorbed 457

ferrihydrite, Zn-humic acid complexes (Zn-HA) at high and low Zn loading, and 458

aqueous Zn2+ as a proxy for outer sphere complexes. Franklinite is easily identified 459

by the high amplitude and multiple frequencies of its spectrum. The spectra for Zn-460

kerolite and Zn/Al hydrotalcite present some similarities, which makes their 461

distinction difficult in a mixture (Panfili et al., 2005). In the LCFs, these two 462

compounds and Zn-sorbed hectorite were grouped as "Zn-phyllosilicate”. Similarly, 463

20

the spectra for Zn-HA at low Zn loading and Zn-sorbed ferrihydrite look similar 464

because Zn is 4-fold coordinated to oxygen atoms and the second shell contribution 465

weak in the two references. Zn-HA at low Zn loading is a proxy for strongly bound 466

inner sphere Zn-organic complexes in tetrahedral configuration (Sarret et al., 1997). 467

In the LCFs, the two tetrahedral species were grouped as "tetrahedral Zn-HA and/or 468

Zn-sorbed ferrihydrite”. Another pair of similar spectra is Zn-HA at high Zn loading 469

and aqueous Zn2+ Because Zn is octahedrally coordinated, and the second shell 470

contribution either weak (Zn-HA) or absent (aqueous Zn). These species are 471

considered as representatives for less-strongly to weakly bound inner sphere Zn-472

organic complexes and outer sphere organic and inorganic complexes (Sarret et al., 473

1997). In the LCFs, these species were grouped as "weakly bound octahedral Zn”. 474

Figure 3a shows the EXAFS spectra for the untreated soil from Mortagne 475

and the two residues, and their reconstructions with four component spectra. In the 476

untreated soil (MRef), Zn is distributed as 60 ± 10 % weakly bound octahedral Zn 477

complexes, 17 ± 10 % Zn-phyllosilicate, 16 ± 10 % tetrahedral Zn-HA and/or Zn-478

sorbed ferrihydrite and 7% franklinite (Figure 3b). The detection limit for this last 479

species is less than 10% because its spectrum has a high amplitude. Zn-sorbed 480

goethite and Zn-sorbed hematite spectra were tested, but neither of them are 481

component species to the data. In a previous study on the same soil (Manceau et al., 482

2000), Zn- phyllosilicate, Zn-sorbed birnessite and Zn sorbed on iron oxyhydroxides 483

were identified as the main Zn species. Another study on a tilled soil near the 484

Mortagne area concluded to the presence of Zn outer-sphere complexes, Zn-organic 485

matter inner-sphere complexes, Zn/Al-hydrotalcite, Zn-phyllosilicate, and 486

magnetite-franklinite solid solutions (Juillot et al., 2003). The F2 extraction 487

removed 41 % of the soil Zn, and most of the weakly bound octahedral Zn pool. The 488

21

F3 extraction removed 15% of the initial soil Zn, the rest of the weakly bound pool, 489

and some of the tetrahedral pool. The Zn-phyllosilicate and franklinite pools were 490

marginally affected. The occurrence of 10-14% Zn as franklinite in the MF2 and 491

MF3 samples is attested by the sharpening of the second oscillation centered at 6 Å-492

1. In this soil the F2 and F3 extractions are relatively specific, affecting mostly the 493

weakly bound octahedral Zn pool. 494

In this soil the weakly bound octahedral Zn measured in the untreated 495

sample (784 mg Zn kg -1 soil) was identical to the IE exchangeable Zn (Epool1+Epool2, 496

786 mg Zn kg -1 soil) and to the amount of Zn extracted by the three first steps of the 497

SSE (F1+F2+F3, 729 mg Zn kg -1 soil) (Table 7). These observations suggest that 498

weakly bound octahedral Zn is the main source of available Zn in this soil which is 499

consistent with our previous EXAFS and isotopic exchange study (Sarret et al., 500

2004). 501

A different behavior is observed for the Dornach soil (Figure 4). Satisfactory 502

fits were obtained with three components for the untreated soil (DRef) and F3 and 503

F4 residues (DF3 and DF4), and with four components for the F5 residue (DF5). 504

The weakly bound octahedral Zn are predominant in the Dornach soil (57 ± 10 %) 505

followed by Zn-phyllosilicate (27 ± 10 %), and tetrahedral Zn-HA and/or Zn-sorbed 506

ferrihydrite (16 ± 10 %). The proportions of Zn species did not change statistically 507

in DF3 and DF4, which suggests that all species were affected to a similar extent by 508

the F3 and F4 treatments (removal of 19 and 41 % of total soil Zn, respectively). 509

The F5 extraction (removal of 19% total soil Zn) targeted preferentially the "weakly 510

bound octahedral Zn" pool. As a consequence, Zn-phyllosilicate was the major 511

species in the DF5 residue. The DF5 spectrum was simulated with Zn-hectorite, 512

whereas DRef, DF3 and DF4 were simulated with Zn-kerolite and/or Zn/Al 513

22

hydrotalcite. Indeed, the third oscillation of the DF5 and Zn-hectorite spectra have 514

the same shape, whereas the shoulder between 7.0 and 7.5 Å-1 in DRef corresponds 515

to a high amplitude feature in Zn-kerolite and/or Zn/Al hydrotalcite spectra (Figures 516

2 and 4a). Therefore, the local environment of Zn seems to have evolved during the 517

selective sequential extraction. A finer description of the nature and structure of 518

these species would require studying the < 2 µm or < 0.2 µm soil fractions by 519

polarized EXAFS (Manceau et al., 2000). A new species is detected in DF5, Zn-520

sorbed birnessite. This species was probably present in the previous samples 521

including those from Mortagne as shown by micro-EXAFS (Manceau et al., 2000), 522

but as a minor component masked by the predominant species. Chemical extractions 523

lacked selectivity in the Dornach soil because the fractional amount of the major 524

species remained unchanged in DF3 and DF4 despite a removal of 19 and 41% of 525

the soil Zn. The weakly bound octahedral Zn complexes which were extracted 526

completely after the third extraction step in Mortagne soil remained predominant in 527

DF3 and DF4. 528

In Dornach the weakly bound octahedral Zn measured in the untreated 529

sample (962 mg Zn kg -1 soil) was similar to the Zn extracted by the four first steps 530

of the SSE (F1+F2+F3+F4, 1023 mg Zn kg -1 soil), but was much higher than the IE 531

exchangeable Zn (Epool1+Epool2, 549 mg Zn kg -1 soil), which itself was also higher 532

than the amount of Zn extracted by the three first steps of the SSE (F1+F2+F3, 326 533

mg Zn kg -1 soil) (Table 7). Although Dornach exhibited a high proportion of 534

weakly bound octahedral Zn, only a small proportion of it was extracted by the 3 535

first extractions of the SSE (22%). We suggest that the IE Zn and the Zn extracted 536

by the 3 first steps of the SSE was indeed present as weakly bound octahedral Zn, 537

but a large fraction of these so-called weakly bound species were neither IE 538

23

exchangeable nor extractable by the three first steps of the SSE. The different 539

behavior of Zn in Dornach and Mortagne may be explained by the difference in soil 540

pH (6.7 for Dornach and 5.1 for Mortagne) and in soil organic matter content (11% 541

for Dornach and 1% for Mortagne) because inner-sphere mineral surface complexes 542

and organically-bound cationic species are more strongly retained at near neutral 543

than at acidic pH. Altogether these results suggest that isotopically exchangeable Zn 544

and therefore available Zn is present as weakly bound octahedral Zn species but that 545

the proportion of weakly bound octahedral Zn that can exchange with Zn2+ in the 546

solution decreases when soil pH and organic matter content increase. 547

548

Conclusion 549

The combination of techniques used in this work (isotopic exchange kinetics, pot 550

experiment with T. caerulescens on soil labeled with 65Zn, selective sequential 551

extraction carried out on 65Zn labeled soils, and EXAFS spectroscopy) gave 552

comprehensive information on the forms and availability of Zn in these heavy metal 553

polluted soils. The main results of this study are summarized in the Table 7. 554

Our results allowed quantifying the number of compartments containing IE 555

Zn in these soils. Three pools were derived from this analysis, the amount of Zn 556

exchangeable within 1 minute (1st pool), the amount of Zn exchangeable between 1 557

minute and apparent isotopic equilibrium (2nd pool) and the amount of Zn that could 558

not be exchanged during the IE kinetic experiment (3rd pool). The experiment 559

conducted with T. caerulescens confirmed that the amount of IE Zn measured in pot 560

experiments was similar to the sum of Zn content present in the 1st and 2nd pools, i.e. 561

that this plant had only access the IE forms of Zn. 562

24

The use of the SSE on 65Zn labeled soils showed that the 3 first extractions 563

(F1, F2 and F3) solubilized the amount of Zn present in the 1st pool and a slight 564

fraction of the Zn present in the 2nd pool. The three last fractions of the SSE (F5, F6 565

and F7) solubilized the Zn from the 3rd pool. We deduced from these observations 566

that the 4th extraction of the SSE solubilized the Zn from the 2nd pool and a fraction 567

of the Zn from the 3rd pool. 568

Finally, EXFAS spectroscopy showed that the 3 first extractions of the SSE 569

solubilized all the weakly bound octahedral Zn in the Mortagne soil, i.e. that the 1st 570

pool was dominated by these Zn species. In Dornach the 5 first extractions of the 571

SSE were necessary to solubilize the weakly bound octahedral Zn. We suggest that 572

in Dornach most of the weakly bound octahedral Zn was not IE nor extractable in 573

the 3 first steps of the SSE. The difference between Mortagne and Dornach could 574

be explained by the highest pH and soil organic matter content of the latter. 575

Altogether these results suggest that isotopically exchangeable Zn and therefore 576

available Zn is present as weakly bound octahedral Zn species but that the 577

proportion of weakly bound octahedral Zn that can exchange with Zn2+ in the 578

solution decreases when soil pH and organic matter content increase. 579

580

Acknowledgements 581

We thank R. Kretzschmar (ETH Zürich) for providing samples of the Dornach and 582

Evin soils, F. van Oort (INRA Versailles) for contributing the Mortagne soil, T. 583

Rösch for her measurements on the ICP, E. Grieder (EAWAG, Dübendorf) for his 584

many hours of gamma measurements, M. Lanson for the preparation of Zn-HA 585

complexes, N. Geoffroy for his help during the EXAFS measurements, the kind 586

support of the staff at beam line ID-26, ESRF (Grenoble) and two anonymous 587

25

reviewers for their constructive remarks. This study was made possible by beam 588

time granted through the European Synchrotron Radiation Facility (ESRF) and 589

funding from the research commission of the ETH, Zürich (TH project n° 8086). 590

591

References 592

Ahnstrom, Z.A.S. & Parker, D.R. 2001. Cadmium reactivity in metal contaminated-593

soils using a coupled stable isotope dilution-sequential extraction procedure. 594

Environmental Science & Technology, 35, 121-126. 595

Almås, Å.R., Salbu, B. & Singh, B.R. 2000. Changes in partitioning of cadmium-596

109 and zinc-65 in soil as affected by organic matter addition and 597

temperature. Soil Science Society of America Journal, 64, 1951-1958. 598

Almås, Å.R., Singh, B.R. & Salbu, B. 1999. Mobility of cadmium-109 and zinc-65 599

in soil influenced by equilibration time, temperature, and organic matter. 600

Journal of Environmental Quality, 28, 1742-1750. 601

Arias, M., Pérez-Novo, C., Osorio, F., Lopez, E. & Soto, B. 2005. Adsorption and 602

desorption of copper and zinc in the surface layer of acid soils. Geoderma, 603

288, 21-29. 604

Ayoub, A.S., McGaw, B.A., Shand, C.A. & Midwood, A.J. 2003. Phytoavailability 605

of Cd and Zn in soil estimated by stable isotope exchange and chemical 606

extraction. Plant and Soil, 252, 291-300. 607

Cardellicchio, N., Cavalli, S., Ragone, P. & Riviello, J.M. 1999. New strategies for 608

determination of transition metals by complexation ion-exchange 609

chromatography and post column reaction. Journal of Chromatography A, 610

847, 251-259. 611

26

Chapman, H. D. & Pratt, P. F. 1961. Methods of analysis for soils, plants, and 612

waters, pp. 169-170. University of California, Division of Agriculture 613

Sciences, Riverside, CA, USA. 614

Cobelli, C., Foster, D. & Toffolo, G. 2000. Tracer Kinetics in Biomedical Research 615

from Data to Model. Kluwer Academic, Plenum Publishers, New York. 616

Degryse, F., Broos, K., Smolders, E. & Merckx, R. 2003. Soil solution 617

concentration of Cd and Zn can be predicted with a CaCl2 soil extract. 618

European Journal of Soil Science, 54, 149-157. 619

FAL, RAC & FAW. 1996. Schweizerische Referenzmethoden der Eidgenössischen 620

Forschungsanstalten, Band 1-4. Eidgenössischen Forschungsanstalten FAL, 621

RAC and FAW, Zürich, Changins, Wädenswil, Switzerland 622

Fardeau, J.C. 1993. Le phosphore assimilable des sols: sa représentation par un 623

modèle fonctionnel à plusieurs compartiments. Agronomie, 13, 317-331. 624

Geiger, G., Federer, P. & Sticher, H. 1993. Reclamation of Heavy-Metal 625

Contaminated Soils - Field Studies and Germination Experiments. Journal of 626

Environmental Quality, 22, 201-207. 627

Juillot, F., Morin, G., Ildefonse, P., Trainor, T.P., Benedetti, M., Galoisy, L., Calas, 628

G. & Brown, G.E. 2003. Occurrence of Zn/Al hydrotalcite in smelter-629

impacted soils from northern France: Evidence from EXAFS spectroscopy 630

and chemical extractions. American Mineralogist, 88, 509-526. 631

Laboratoire de Spectrochimie Infrarouge et Raman (LASIR). 2000. Rapport 632

d'activités 1997-2000. Université des Sciences et Technologies de Lille, 633

Villeneuve d’Ascq, France p57. 634

27

Loeppert, R.L. & Inskeep, W.P. 1996. Iron. In: Methods of soil analysis. Part 3. 635

Chemical Methods (ed. P.J. Sparks), pp. 639-664. SSSA Book Series: 5. 636

SSSA, Madison, Wisconsin. 637

Lorentzen, E.M.L. & Kingston, H.M. 1996. Comparison of microwave-assisted and 638

conventional leaching using EPA method 3050B. Analytical Chemistry, 68, 639

4316-4320. 640

Manceau, A., Marcus, M. A. & Tamura N. 2002. Quantitative speciation of heavy 641

metals in soils and sediments by synchrotron X-ray techniques. In 642

Applications of Synchrotron Radiation in Low-Temperature Geochemistry 643

and Environmental Science, Vol. 49 (eds. P. Fenter, M. Rivers, N. Sturchio, 644

& S. Sutton), pp. 341-428. Reviews in Mineralogy and Geochemistry, 645

Mineralogical Society of America. Washington DC. 646

Manceau, A., Lanson, B., Schlegel, M.L., Harge, J.C., Musso, M., Eybert-Berard, 647

L., Hazemann, J.L., Chateigner, D. & Lamble, G.M. 2000. Quantitative Zn 648

speciation in smelter-contaminated soils by EXAFS spectroscopy. American 649

Journal of Science, 300, 289-343. 650

Manceau, A., Tamura, N., Celestre, R.S., MacDowell, A. A., Geoffroy, N., Sposito, 651

G. & Padmore, H.A. 2003. Molecular-scale speciation of Zn and Ni in soil 652

ferromanganese nodules from loess soils of the Mississippi basin. 653

Environmental Science & Technology, 37, 75-80. 654

Nolan, A.L., Zhang, H. & McLaughlin, M.J. 2005. Prediction of Zinc, Cadmium, 655

Lead, and Copper availability to wheat in contaminated soils using chemical 656

speciation, diffusive gradients in thin films, extraction, and isotopic dilution 657

techniques. Journal of Environmental Quality, 34, 496-507. 658

28

Panfili, F.R., Manceau, A., Sarret, G., Spadini, L., Kirpichtchikova, T., Bert, V., 659

Laboudigue, A., Marcus, M.A., Ahamdach, N. & Libert, M.F. 2005. The 660

effect of phytostabilization on Zn speciation in a dredged contaminated 661

sediment using scanning electron microscopy, X-ray fluorescence, EXAFS 662

spectroscopy, and principal components analysis. Geochimica et 663

Cosmochimica Acta, 69 (9): 2265-2284 664

Ressler, T. 2000. WinXAS version 2.0. Livermore, CA, USA. 665

Roberts, D.R., Scheinost, A.C. & Sparks, D.L. 2002. Zinc speciation in a smelter-666

contaminated soil profile using bulk and microspectroscopic techniques. 667

Environmental Science & Technology, 36, 1742-1750. 668

Salbu, B., Krekling, T. & Oughton, D.H. 1998. Characterisation of radioactive 669

particles in the environment. Analyst, 123, 843-849. 670

Sarret, G., Balesdent, J., Bouziri, L., Garnier, J., Marcus, M., Geoffroy, N., Panfili, 671

F. & Manceau, A. 2004. Zn speciation in the organic horizon of a 672

contaminated soil by micro X-ray fluorescence, micro and powder EXAFS 673

spectroscopy and isotopic dilution. Environmental Science & Technology, 674

38, 2792 -2801. 675

Sarret, G., Manceau, A., Hazemann, J.L., Gomez A. & Mench, M. 1997. EXAFS 676

study of the nature of zinc complexation sites in humic substances as a 677

function of Zn concentration. Journal de Physique iv, 7, 799-802. 678

Scheinost, A.C., Kretzschmar, R. & Pfister, S. 2002. Combining selective sequential 679

extractions, x-ray absorption spectroscopy, and principal component analysis 680

for quantitative zinc speciation in soil. Environmental Science & 681

Technology, 36, 5021-5028. 682

29

Schlegel, M. L., Manceau, A., Charlet, L., Chateigner D. & Hazemann, J. L. 2001. 683

Sorption of metal ions on clay minerals. III. Nucleation and epitaxial growth 684

of Zn phyllosilicate on the edges of hectorite. Geochimica Et Cosmochimica 685

Acta 65 (22), 4155-4170. 686

Schlegel, M.L. & Manceau, A. 2006. Evidence for the nucleation and epitaxial 687

growth of Zn phyllosilicate on montmorillonite. Geochimica Et 688

Cosmochimica Acta, 70, 901-917. 689

Siegenthaler, A., Häni, H. & Stauffer, W. 1999. Effect of high sewage sludge and 690

pig slurry application. Agrarforschung, 6(5), 1-8. 691

Sinaj, S., Mächler, F. & Frossard, E. 1999. Assessment of isotopically exchangeable 692

zinc in polluted and nonpolluted soils. Soil Science Society of America 693

Journal, 63, 1618-1625. 694

Sinaj, S., Dubois, A. & Frossard, E. 2004. Soil isotopically exchangeable zinc: a 695

comparison between E and L values. Plant and Soil, 261, 17-28. 696

Smith, F. W. 1981. Availability of soil phosphate to tropical pasture species. In: 697

Proceedings of XIV International Grassland Congress Lexington, KY USA 698

(ed. by A. J. Smith & V. W. Hays). Westview press, Bolder, CO USA. 699

Thiry, M., Huet-Taillanter, S. & Schmitt, J.M. 2002. The industrial waste land of 700

Mortagne-du-Nord (59) - I - Assesment, composition of the slags, 701

hydrochemistry, hydrology and estimate of the outfluxes. Bulletin de la 702

Société Géologique de France, 173, 369-381. 703

Weissenhorn, I., Mench, M. & Leyval, C. 1995. Bioavailability of Heavy-Metals 704

and Arbuscular Mycorrhiza in a Sewage-Sludge-Amended Sandy Soil. Soil 705

Biology & Biochemistry, 27, 287-296. 706

30

Young, S.D., Tye, A., Carstensen, A., Resende, L. & Crout, N. 2000. Methods for 707

determining labile cadmium and zinc in soil. European Journal of Soil 708

Science, 51 (1), 129-136 709

Young, S.D., Zhang, H., Tye, A.M., Maxted, A., Thums, C. & Thorton, I. 2006. 710

Characterizing the availability of metals in contaminated soils. I. The solid 711

phase: sequential extraction and isotopic dilution. Soil Use and Management, 712

21, 450-458 713

31

Figure Captions 714

Figure 1 Change in the fraction of radioactivity remaining in solution (rt/R) during 715

14 days in six soils that have been polluted with heavy metals (1a INRA soil, 1b 716

IUL SS soil, 1c IUL PS soil, 1d Evin soil, 1e Dornach soil, 1f Mortagne soil). The 717

points represent the experimental data, the solid line the values predicted from the 718

model developed for each soil, and the dotted lines the 80% confidence limits of the 719

model. 720

721

Figure 2 Zn K-edge EXAFS spectra (χ(k) k3, with k: wave number) of reference Zn 722

compounds used to model the spectra for the soils and extraction residues: 723

Franklinite (ZnFe2O4), Zn-sorbed birnessite (adsorption at pH 4, Zn/Mn = 0.134) 724

(Manceau et al., 2000), Zn-substituted phyllosilicate ((Zn, Mg) kerolite 725

Zn2.1Mg0.9Si4O10(OH)2, nH2O, Schlegel and Manceau, 2006 and Zn-substituted 726

hectorite, Schlegel et al., 2001), Zn/Al hydrotalcite (Zn2Al(OH)6(CO3)0.5, n H2O), 727

Zn-sorbed ferrihydrite containing 1500 mg kg-1 Zn (Manceau et al., 2000), Zn-728

humic acid complexes at low (600 mg kg-1 Zn) and high (3.2 % Zn) Zn loading 729

(Sarret et al., 1997), and aqueous Zn (pH 4). 730

731

Figure 3 a) Zn K-edge EXAFS spectra (solid line) and linear combination fit 732

(dashed line) for the Mortagne untreated reference soil (MRef), for the residue left 733

after the second (MF2) and third (MF3) extraction of the SSE. b) Distribution of Zn 734

species derived from the fits. The error bars correspond to 10% of total Zn content. 735

736

Figure 4 a) Zn K-edge EXAFS spectra (solid line) and linear combination fits 737

(dashed line) for the Dornach untreated reference soil (DRef), for the residue left 738

32

after the third (DF3), fourth (DF4) and fifth (DF5) extraction of the SSE. b) 739

Distribution of Zn species derived from the fits. The error bars correspond to 10% of 740

total Zn content. 741

33

Figure 1 742

0.01

0.1

10 5000 10000 15000 20000 25000

minutes

r t /R

Fig 1.a INRA soil

743

0.001

0.01

0.1

10 5000 10000 15000 20000 25000

minutes

r t /R

Fig 1.b IUL SS soil

744

0.1

10 5000 10000 15000 20000 25000

minutes

r t /R

Fig 1.c IUL PS soil

745

34

0.1

10 5000 10000 15000 20000 25000

minutes

r t /R Fig 1.d Evin soil

746

0.001

0.01

0.1

10 5000 10000 15000 20000 25000

minutes

r t /R

Fig 1.e Dornach soil 747

0.1

10 5000 10000 15000 20000 25000

minutes

r t /R

Fig 1.f Mortagne soil 748

35

Figure 2 749

750

36

Figure 3 751

0 500 1000 1500 2000

MRef

MF2

MF3

Zn (mg kg-1 soil)

Franklinite

Zn-phyllosilicates

Weakly bound octahedral ZncomplexesTetrahedral Zn-HA and/or Zn-sorbed ferrihydrite

7% 60% 16%

25%

32% 33%

42%

44%14%

17%

10%

a b

37

Figure 4 752

753

754

755

0 500 1000 1500 2000

Dref

DF3

DF4

DF5

Zn (mg kg-1 soil)

Zn-phyllosilicates

Weakly bound octahedralZn complexes

Tetrahedral Zn-HA and/orZn-sorbed ferrihydrite

Zn-sorbed birnessite

27% 57% 16%

28% 55% 17%

20% 66% 14%

52% 29%

10%

11%

a b

38

Table 1 Selected characteristics of the six soils polluted with heavy metals 756 757

Characteristic INRA IUL SS IUL PS Evin Dornach Mortagne

Soil type Eutric Fluvisol Orthic Luvisol Orthic Luvisol Gleyic Luvisol Calcaric Regosol Dystric Cambisol Pollutant source Sewage Sludge Sewage Sludge Pig Slurry Pb/Zn Smelter Cu/Ni/Zn Smelter Zn Smelter Land use arable Arable Arable forest grassland grassland Sampling depth / cm 0-20 0-20 0-20 2-5 2-5 5-40 Sand / %a 80 57 58 28 15 66 Clay / %a 7 15 15 20 37 7 Organic matter / %b 2 3 2 4 11 1 CaCO3 / %c n.d. n.d. n.d. n.d. 5 n.d. Oxalate extractable Fe / g kg-1 soild 2.14 11.11 5.11 3.17 2.70 1.88 EDTA extractable Zn / mg kg-1 soile 166 15 9 435 349 406 pHf 5.7 6.0 4.6 5.0 6.7 5.1 CEC / mmolc kg-1g 101 164 161 185 432 67 Base saturation / %g 47.8 17.5 45.2 85.6 42.2 18.8 Total Zn / mg kg-1 soili 680 (5.4) 162 (0.5) 87.9 (1.7) 1647 (3.1) 1687 (3.1) 1307 (8.7) n.d. non detectable a Soil texture was measured via sedimentation with (NaPO4)6 as a dispersion agent (FAL, RAC & FAW, 1996). b Organic matter was measured by titration (FAL, RAC & FAW, 1996). c CaCO3 was quantified using concentrated HCl (FAL, RAC & FAW, 1996). d Oxalate extractable Fe was determined according to Loeppert and Inskeep (1996). e EDTA-NH4Ac extractions were performed using 50ml of extractant added to 10g dry soil (FAL, RAC & FAW, 1996). f pH was measured using a 1:2.5 soil solution ratio of 0.01M CaCl2 after 24 hours of gentle shaking. g CEC and base saturation were determined using BaCl2 method (FAL, RAC & FAW, 1996). i Total Zn obtained after direct digestion (n=3) and standard error (in parentheses). 758

39

Table 2 Selective sequential extraction used for the fractionation of Zn based on Salbu et al. (1998) and Lorentzen and Kingston (1996) 759

760

Extraction step Reagents Procedure Proposed binding mechanism targeted a

F1 H2O 1 hours at 20° C Water soluble

F2 1 M NH4OAc, pH 7 2 hours at 20° C Reversible physisorption

F3 1 M NH4OAc, pH 5 (HNO3) 2 hours at 20° C Reversible electrosorption

F4 0.04 M NH2OH-HCl in 4.4 M CH3COOH 6hours at 80° C Irreversible chemisorption / reduction

9.7 M H2O2, pH 2 (HNO3) 5.5 hours at 80° C

F5 followed by Chemisorption / oxidation

3.2 M NH4OAc in 4.4 M HNO3 30 minutes at 20° C

F6 7 M HNO3 6 hours at 20° C Chemisorption / crystalline

Total Digestion

F7 HNO3 , H2O2 and HCl Open-Microwave Digestion 45 minutes at 70°, 90° C Residue

a Proposed by Salbu et al. (1998)

40

Table 3 Average Zn concentration in dilute CaCl2 extracts (CZn) measured during the isotopic exchange experiments and amounts of Zn 761

exchangeable within 1 minute (Epool1), moderately isotopically exchangeable Zn (Epool2) and slowly or not exchangeable Zn (Epool3) 762

calculated for six soils polluted with heavy metals. The average data is followed by the standard error (se) given between parentheses. 763

Soil CZn Epool1 Epool2 Epool3 Average Se Average se Average se Average se

_____ /mg Zn l-1 ______ _______________________ /mg Zn kg-1soil _______________________

INRA 0.50 0.030 50.7 (3.0) 167 (5.9) 463 (8.8) IUL SS 0.02 0.003 5.7 (0.6) 28.6 (2.5) 128 (3.1) IUL PS 0.38 0.006 10.5 (0.3) 12.6 (0.9) 64.8 (0.7) Evin 12.4 0.149 697 (11.6) 305 (14.4) 645 (10.6) Dornach 0.15 0.007 181 (1.9) 368 (18.5) 1138 (18.6) Mortagne 29.9 0.429 612 (10.2) 174 (4.1) 521 (7.0)

764

41

Table 4 Average values for plant aerial dry matter, Zn content of aerial parts and L value measured with Thlaspi caerulescens in six soils 765

polluted with heavy metals. The standard errors (se) are given between parentheses. 766

767

Soil Yield Plant Zn content L value Average se Average se Average se / g DM kg-1 soil / g kg-1 DM / mg Zn kg-1 soil INRA 16.5 (0.5) 0.4 (0.1) 224 (14.0) IUL SS 23.4 (0.3) 1.2 (0.1) 36 (0.2) IUL PS 20.6 (1.1) 1 (0.1) 22.6 (0.7) Evin 21.7 (0.4) 4 (0.5) 828 (14.1) Dornach 20.4 (0.8) 2.2 (0.5) 564 (0.8) Mortagne 21.7 (0.7) 5.9 (0.3) 737 (4.7)

42

Table 5 Average amount of Zn recovered in the different fractions of the selective sequential extraction after 20, 85 and 120 days of 768

incubation in six soils polluted by heavy metals. The results are expressed in mg Zn kg-1 soil and the standard errors (se) are given between 769

parentheses. 770

771

Fractions INRA IUL SS IUL PS Evin Dornach Mortagne Average se Average se Average se Average se Average se Average se

_____________________________________________________________________ /mg Zn kg-1soil _____________________________________________________________________ F1 2 (0.5) 0.2 (0.1) 1.2 (0.5) 23.7 (6.8) 1.2 (0.5) 28.1 (3.8) F2 26.7 (1.8) 1.8 (0.2) 2.5 (0.2) 499 (27.9) 78.7 (4.2) 505 (26.0) F3 103 (11.7) 10.4 (1.6) 8 (0.9) 248 (2.9) 246 (23.4) 195 (4.8) F4 217 (10.0) 55.1 (3.5) 19.8 (1.0) 362 (22.1) 696 (44.8) 285 (20.1) F5 71.6 (21.5) 24.4 (2.3) 11.2 (1.6) 256 (18.1) 317 (48.9) 64 (11.0) F6 279 (38.6) 74.9 (10.6) 43.6 (4.9) 256 (36.8) 165 (37.6) 101 (11.9) F7 10.1 (2.2) 21.4 (2.7) 12.0 (2.9) 45.0 (10.9) 28.9 (5.1) 26.2 (4.8) Total extracted 709 (19.5) 188 (12.6) 98.2 (6.0) 1690 (42.3) 1533 (64.5) 1205 (51.2)

772

773

774

43

Table 6 Mean specific activities calculated for each fraction of the selective sequential extraction over 120 days for 6 soils polluted with 775

heavy metals. The values for each fraction are normalized by the 65Zn introduced and the total soil Zn. Standard errors (se) are given 776

between parentheses. 777

778

Fraction INRA IUL SS IUL PS Evin Dornach Mortagne Average se Average se Average se Average se Average se Average se F1 1.95 (0.21) 1.15 (0.16) 2.42 (0.86) 1.32 (0.38) 1.22 (0.43) 1.10 (0.16) F2 2.51 (0.05) 4.24 (0.12) 3.54 (0.28) 1.53 (0.12) 2.10 (0.29) 1.25 (0.05) F3 2.26 (0.11) 3.48 (0.26) 3.10 (0.17) 1.65 (0.10) 1.80 (0.06) 1.27 (0.05) F4 1.67 (0.10) 1.89 (0.19) 2.27 (0.17) 1.00 (0.11) 0.70 (0.04) 0.58 (0.08) F5 0.49 (0.06) 0.38 (0.01) 0.47 (0.02) 0.37 (0.06) 0.42 (0.06) 0.17 (0.03) F6 0.01 (0.01) 0.07 (0.02) 0.08 (0.02) 0.12 (0.03) 0.22 (0.08) 0.04 (0.01) F7 0.01 (0.01) 0.02 (0.01) 0.02 (0.01) 0.04 (0.02) 0.08 (0.02) 0.01 0.00)

779

44

Table 7 Comparison of the data obtained by the various techniques 780 781 Soil INRA IUL-SS IUL-PS Evin Dornach Mortagne Total Zn / mg Zn kg soil-1 680 162 87.9 1647 1687 1307

pH 5.7 6.0 4.6 5.0 6.7 5.1 Concentration of Zn in CaCl2; CZn / mg Zn l-1

0.50 0.01 0.38 12.3 0.15 29.9

Epool1/ Epool2 /Epool3 / mg Zn kg soil-1 50.7 / 167 / 463 5.7 / 28.6 / 128 10.5 / 12.6 / 64.8 697 / 305 / 645

181 / 368 / 1138

612 / 174 / 521

L values / mg Zn kg soil-1 224 36 22.6 828 564 737

Extraction steps, by decreasing order of Zn recovery

F6>F4>F3> F5>F2>F7>F1

F6>F4>F5> F7>F3>F2>F1

F6>F4>F7> F5>F3>F2>F1

F2>F4>F6-F5>

F3>F7>F1 F4>F5>F3> F6>F2>F7>F1 F2>F4>F3> F6>F5>F1>F7

Distribution of Zn species in the untreated soil / % a

57% weakly bound octahedral Zn + 27% Zn- phyllosilicates + 16% tetrahedral Zn-HA and/or Zn-sorbed ferrihydrite

60% weakly bound octahedral Zn + 17% Zn- phyllosilicates + 16% tetrahedral Zn-HA and/or Zn-sorbed ferrihydrite + 7% franklinite

Comparison of results obtained by the different methods / mg Zn kg soil-1

F1+F2+F3+F4 (1023) = weakly bound octahedral Zn (962) > Epool1+ Epool2 (549) > F1+F2+F3 (326)

F1+F2+F3+F4 (1014) > weakly bound octahedral Zn (784) = Epool1+ Epool2 (786) = F1+F2+F3 (729)

a Determined only for Dornach and Mortagne soils. 782 783